The rising incidence of melanoma and the lack of effective treatments for advanced disease represents an important public health problem. Recent studies have demonstrated that melanomas arising from skin on the palms and soles (acral melanomas), on mucosal membranes, and from skin with chronic sun-induced damage (CSD) have distinctive patterns of chromosomal alterations as compared with the more common subtypes of melanoma arising on skin without CSD. Unlike melanoma occurring on skin without CSD, these less common subtypes of melanoma are characterized by increased copy numbers or amplifications of the chromosome 4q12 locus, with an associated overexpression of the resident tyrosine kinase protein c-KIT and/or mutations of c-KIT. Among the several mutations of c-KIT identified in these melanomas are the activating mutations that, in other tumor types such as gastrointestinal stromal tumors (GIST), have been associated with sensitivity to treatment with small molecule inhibitors of c-KIT. Imatinib mesylate may offer this particular subset of melanoma patients a new therapeutic option. This application proposes to test the hypothesis that, in tumors with somatic alterations of c-KIT, imatinib mesylate will achieve an objective response rate of at least 30 percent. A Simon two-stage minimax study design will be utilized. A response rate of 10 percent or less will not be considered promising, and a 30 percent response rate will be considered promising. The probabilities of a type I and type II error are both set at 0.10. In the initial stage, 16 patients will be enrolled. If fewer than 2 patients achieve a response, the trial will be closed and imatinib mesylate deemed ineffective in this patient population. If 2 or more responses are observed, 9 additional patients will be accrued until a total of 25 evaluable patients have been accrued. If 5 or more responses are observed in the 25 patients, then imatinib mesylate will be considered worthy of further testing. This design yields at least a 90 percent probability of a positive result if the true response rate is greater than or equal to 30 percent. Patients will be enrolled from 3 local sites. All tumors will be screened using florescence in-situ hybridization (FISH) and DNA sequencing. Only patients with tumors harboring c-KIT amplifications or juxtamembrane domain mutations (the mutation site associated with the greatest sensitivity to imatinib mesylate in GIST) will be eligible. In addition, correlative studies will be performed on each patient's tumor including immunohistochemistry for CD117 as well as comparative genomic hybridization (CGH) to further assess for amplification at 4q12. If all eligibility criteria are met, patients will be prescribed 100 milligrams of imatinib mesylate capsules and will take 4 capsules twice a day (400 mg BID) by mouth on a continual basis. Imaging studies will be performed after 6 weeks of therapy. Patients who show a response or stable disease and have no evidence of progression will continue receiving daily imatinib mesylate. Dose reductions are allowed in the setting of toxicity. Imaging studies will be performed on an every 6 week schedule.